Tuesday on MSNBC, Lester Holt interviewed reporter Roland Jones about tradesports.com, a website on which people have been betting real money on the presidential election. The goings on at the Dublin-based company, as well as similar operations like the Iowa Electronics Futures Market, have gotten plenty of attention over the course of the campaign, with some media outlets suggesting they’re more reliable than polls.
Jones echoed the conventional wisdom on the matter Tuesday when he said, “The thinking goes that…if a pollster asks you who you think will win the election, you have no incentive to tell the truth. But if you’re betting your own hard-earned money, you’re going to put a lot of research into it, and so the thinking is that the online betting sites, their odds are likely to be more accurate.”
That’s a persuasive argument on its face, and it explains why outlets from the Economist to MSNBC to the New York Times have taken Tradesports seriously. But while there’s value — or at least entertainment — for political junkies in the betting at Tradesports and its brethren, many in the media have been a little too quick to treat such sites as windows into the electorate.
For starters, Tradesports bettors are a self-selecting bunch — political junkies with disposable income — and we don’t know to what degree they’re betting rationally and to what degree they are, in the words of University of Nevada professor Bill Thompson, “simply cheering on their side.” (There’s also reason to assume that investors in a stock-like market have business backgrounds and, as such, perhaps, tilt more right than left.) In addition, bettors may be pooling their investments, and, if bloggers are to be believed, they may even be trying to manipulate re-election futures. The sample size is small as well; as Jesse Taylor points out, Tradesports so far has trafficked in about $13 million in presidential contracts. By way of context, $126 million in IBM stock was traded between 9:30 to 10:45 A.M. yesterday.
That isn’t to say that there isn’t some value in looking at Tradesports, which presently shows Bush leading Kerry 61-40. Its fluctuations following the release of September’s jobs report gave an immediate sense of the potential impact of the news, and the markets were closer to accurate than some opinion polls in predicting the outcome of Australian elections. Further, the director of the Iowa Electronic Futures market says that between 1988 and 2000, three times out of four the Iowa market was closer than polls to final election results.
But as what is universally acknowledged to be a tight race winds down to its final days, there’s something a little fishy about any indicator that gives either candidate a 20-point lead. Just like polls, markets like Tradesports are tools of limited predictive value, no matter how much enthusiasm is shown by the talking heads.